The incidence of malaria in highland areas of Africa is growing, with severe epidemics occurring more frequently. Little is known about the environmental and entomological determinants of highland malaria. This study is designed to identify risk factors for disease in Kipsamoite, a highland area in the western part of Kenya. In collaboration with colleagues working there, field observations and statistical analyses will be undertaken to address two specific aims: 1) determine whether specific ecological and household factors modulate malaria risk, and 2) evaluate the spatial and temporal patterns of malaria cases and their relation to environmental variables. Standard epidemiologic methods will be combined with various environmental measurements, new approaches using satellite imagery, and spatial statistical analyses of spatial-temporal patterns. A malaria case-control study design will be used to examine environmental and household parameters hypothesized to affect transmission. In addition to risk aversion behaviors and demographics, house construction and surrounding ecological characteristics will be studied in relation to mosquito breeding sites. It is hypothesized that the distribution of cases in this unstable malaria region is non-random, both spatially and temporally. Using Geographic Information Systems (GIS) and new spatial-temporal statistical methods, patterns of cases in relation to environmental variables will be studied. Because data will be obtained at an unusually fine spatial scale, risk related to individual households and breeding sites may be identified. Results of this "Small Research Grant" investigation should produce the necessary preliminary data to expand the effort and propose a full "R01" application. Hopefully, such findings will not only help to identify additional research hypotheses for critical environmental risk factors affecting malaria in highland African sites, but will also suggest simple and effective control measures that can be easily tested for future application. [unreadable] [unreadable]